Microsoft uses the VCR defense against The New York Times

Forty years ago, Hollywood tried and failed to sue the home VCR out of existence. Now, Microsoft is dusting off that exact cassette tape to defend its multi-billion-dollar AI empire.

Microsoft has officially filed a motion to dismiss three claims in The New York Times’ copyright infringement lawsuit against the tech giant and OpenAI. In its court filing, Microsoft’s legal team argued that copyright law has never stopped technological leaps before, comparing large language models to player pianos, copy machines, personal computers, search engines, and, most notably, the VCR.

The Betamax defense returns

At the center of Microsoft’s argument is the decades-old legal doctrine establishing that a technology cannot be outlawed simply because it can be abused, as long as it has substantial lawful use.

Microsoft pushed back hard against claims that it induced Copilot users into copyright infringement. According to the motion, The Times failed to identify even a single real-world Copilot user who engaged in direct infringement. Microsoft argued that holding an AI developer liable under contributory infringement without direct user proof would undermine the legal precedent that protected consumer electronics forty years ago.

Microsoft also attacked The Times' claim under the Digital Millennium Copyright Act (DMCA), which alleged that the companies systematically stripped copyright management information, such as author bylines, from training data. Microsoft reminded the court that judges had already thrown out similar DMCA claims in other high-profile AI lawsuits, including the one brought by author Sarah Silverman.

A flawed analogy?

Counsel for The Times was not amused by the retro tech nostalgia.

Ian Crosby, partner at Susman Godfrey and lead attorney for the publisher, pointed out that Microsoft never actually denied copying the stories. He noted that the comparison is bizarre because VCR makers never claimed they needed to commit massive copyright infringement just to build the device in the first place.

This follows OpenAI’s own motion to dismiss, in which the startup claimed that The Times essentially hacked and tricked ChatGPT into spitting out verbatim articles using unrealistic prompts. Both companies are now working overtime to frame AI as a neutral tool rather than an automated content scraper.

Why it matters

This legal battle could define the economic foundation of generative AI for the next decade.

From Microsoft’s perspective, invoking the VCR defense is an aggressive and clever tactic to shift liability onto end users. If courts view Copilot as just another photocopier, tech companies can wash their hands of what users generate.

However, we believe the analogy has a glaring flaw: a VCR leaves the factory completely empty of copyrighted video, while large language models are fundamentally made of the copyrighted texts used to train them. Conflating the act of building a model with consumer use of a tool is a massive stretch.

If The Times defeats this motion, it could force tech giants to negotiate licensing deals for every scrap of public training data, likely driving up API costs across the entire ecosystem. If Microsoft prevails, open web scraping for AI training will receive a monumental stamp of judicial legitimacy.